Compliance/PCI DSS 4.0

PCI DSS 4.0 Evidence Mapping for AI Agents

PCI DSS 4.0 fully replaced 3.2.1 on March 31, 2024, and the future-dated requirements landed March 31, 2025. If an AI agent in your stack ever touches the Cardholder Data Environment, Requirements 7, 8, and 10 apply to each governed tool call it makes. Runtime authorization is how those requirements get enforced for non-human identities.

Last updated: May 20, 2026

What is PCI DSS 4.0?

The Payment Card Industry Data Security Standard version 4.0 was published by the PCI Security Standards Council on March 31, 2022. PCI DSS v3.2.1 was retired March 31, 2024; future-dated requirements in v4.0 became fully enforceable on March 31, 2025. PCI DSS v4.0.1 was published in June 2024. The standard applies to any organization that stores, processes, or transmits cardholder data (CHD) or sensitive authentication data (SAD). Enforcement is contractual through card brands and acquiring banks; non-compliance and confirmed breaches can create material assessments, remediation costs, and acquiring-bank scrutiny.

Why it applies to AI agents

Agents are increasingly deployed in payment workflows: dispute resolution, refund triage, billing reconciliation, fraud review, customer self-service. Even agents that only read masked PANs are subject to PCI DSS controls if they operate inside the CDE or interact with systems that do. The fact that an agent is non-human does not make the requirements looser. Requirement 7.2.5 calls out application and system accounts as a distinct category that needs explicit access management.

The risk is the same risk that drives every other PCI control: an unauthorized action can expose primary account numbers, magstripe data, or CVV2 codes. An agent with broad access to a payments API and prompt injection in its inputs creates breach exposure if the tool path is not governed.

Control mapping: PCI DSS 4.0 requirements to Veto features

The table maps the requirements most directly engaged by AI agent systems operating in or adjacent to the CDE.

RequirementDescriptionVeto feature
7.2.1Access control model defined for system components in scopePolicy-as-code enumerates allowed tools, arguments, and approval rules per agent identity
7.2.2Access assigned based on job classification and least privilegeRole-based policy files separating support, billing, fraud-review, and reconciliation agents
7.2.4All user accounts reviewed at least every six monthsPolicy review reminders; CODEOWNERS-required reapproval of agent policies on a configurable cadence
7.2.5Application and system accounts managed with intent and minimum privilegePer-agent API keys, scoped to project/org; documented intent in policy comments
8.2.1Unique ID for each user before allowing access to system componentsUnique agent ID and API key per identity; cross-tenant isolation
8.6.1Accounts used by systems/applications managed to prevent unauthorized useKey rotation history; environment-scoped keys; alerting on unusual usage patterns
8.6.2Passwords/passphrases for application/system accounts are not hard-coded in codeAPI keys delivered via secret manager integrations; never embedded in policy YAML
10.2.1Audit logs capture all user access to system components and CHDGoverned CDE tool calls can be recorded with agent ID, tool, arguments (PAN redacted), policy, and outcome
10.2.2Audit logs capture user identification, type of event, timestamp, success/failure, origin, and affected dataDecision record schema covers the core audit fields per record
10.4.1Audit logs reviewed at least daily for security eventsPolicy-violation views and exports that support your daily review workflow and reviewer sign-off
10.5.1Audit log history retained for at least one year, with three months immediately availableConfigurable retention and export paths; align hot storage and archive policy with your PCI evidence plan
12.3.1Critical technologies have approved use cases, authentication, list of devices, acceptable usePolicy YAML serves as the approved-use document; CI gates restrict unapproved tools

Evidence Veto provides

A QSA performing a Report on Compliance (ROC) needs documented evidence per requirement. Self-assessment questionnaires (SAQ-A through SAQ-D) need the same artifacts at lower attestation depth.

Access matrices

Per-agent policy YAML enumerates allowed tools and arguments. Reviewable by QSA as the access control model required by 7.2.1.

Decision records

Decision record entries per Req 10.2.1 and 10.2.2 with PAN redaction configured so decision records do not need to become a CHD repository.

Daily review records

Reviewer sign-off on flagged decisions documents the daily log review required under 10.4.1.

Six-month reviews

Policy review timestamps and CODEOWNERS approvals document the periodic access review required by 7.2.4.

Implementation timeline

March 31, 2022PCI DSS v4.0 published
March 31, 2024PCI DSS v3.2.1 retired; v4.0 becomes the only valid version
June 2024PCI DSS v4.0.1 errata release
March 31, 2025Future-dated requirements (new MFA, scripting, authentication controls) become fully enforceable
AnnuallyROC or SAQ submitted to acquirers; quarterly ASV scans where applicable

Card-brand assessments, acquiring-bank action, forensic reviews, and remediation costs can become material after non-compliance or a confirmed breach.

Frequently asked questions

When did PCI DSS 4.0 become mandatory?
PCI DSS v4.0 was published March 31, 2022. The transition window allowed organizations to continue using v3.2.1 through March 31, 2024. After that date, PCI DSS v4.0, including the v4.0.1 update released in June 2024, is the applicable v4 line. Future-dated requirements, including new MFA, scripting, and authentication controls, became fully effective on March 31, 2025. Assessments after that date need to address the applicable v4.0 or v4.0.1 requirements.
Which PCI DSS requirements apply to AI agents?
Requirements 7 (Restrict Access by Business Need to Know), 8 (Identify Users and Authenticate Access), 10 (Log and Monitor All Access), and 12 (Support Information Security with Organizational Policies) apply most directly. If an agent ever touches cardholder data (CHD) or the Cardholder Data Environment (CDE), each governed tool call is subject to those requirements.
Can AI agents have access to cardholder data?
Yes, but only under the same conditions as every other access vector. Requirement 7.2.1 requires access control models defined for system components in the CDE. Requirement 7.2.5 explicitly covers application and system accounts, which is the closest analogue to AI agent identities. Requirement 8.6 governs application and system account use, requiring documented intent, regular review, and no shared accounts.
What evidence does a QSA expect from an AI agent deployment?
Qualified Security Assessors expect documented access matrices for each agent identity (Req 7.2.1), evidence of unique authentication for each agent (Req 8.2.1, 8.6.1), tool-call decision records for governed CDE actions (Req 10.2.1), daily log review evidence (Req 10.4.1), and retention of decision records for at least one year with three months immediately available (Req 10.5.1). Veto can support the evidence path for these controls when governed CDE actions run through the wrapped decision path.

Related evidence resources

If an agent can touch PAN or the CDE, it needs PCI-grade access and log evidence. Make the boundary reviewable.